M PRA

Munich Personal RePEc Archive

Impact of monetary policy on lending

and deposit rates in Pakistan: Panel data
analysis

Mohsin, Hasan Muhammad

Pakistan Instittute of Development Economics Islamabad
Pakistan

January 2011

Online at http://mpra.ub.uni-muenchen.de/33301/
MPRA Paper No. 33301, posted 10. September 2011 / 14:19
 Impact of Monetary Policy on Lending and Deposit
rates in Pakistan: Panel data analysis

Hasan Muhammad Mohsin, PhD

Research Economist
Pakistan Institute of Development Economics
Quaid-i-Azam University Campus
PO box 1091
Islamabad
Email: hasanmohsin@pide.org.pk
Phone: 92-51-9248080

8/10/2011

The study estimates impact of monetary policy rate on lending and deposit rates in
Pakistan using bank type data from November 2001 to March 2011. The study finds
evidence of long run relationship between lending and discount rate but deposit rate is not
cointegrated with discount rate, monetary policy instrument. The study also finds an
increase in the lending rate pass through rate during contractionary monetary policy times
(2005-2010), whereas deposit rate pass through remains same. The study finds that
overall banks pass on only 20% impact of a change in discount rate to the lenders in first
month implying it is not complete. There is also significant difference in various bank
types pass through rates. The pass through of deposit rate is further low at 0.16 as
revealed by short run analysis. It implies that the effectiveness of monetary policy is
limited in Pakistan.
Impact of Monetary Policy on Lending and Deposit rates in Pakistan: Panel data
analysis

I. Introduction:

Interest rate is one of the tools of monetary policy. Pass through is the transmission of
bench mark interest rate, Discount rate (DR) in our case, to lending and deposit rates in
the economy. The pass through is complete if any change in DR is transmitted to the
lending and deposit rates immediately. 1 The completeness implies monetary policy is
very effective and central bank can influence output and consumption without much
delay.
Many studies have estimated the degree of pass through for the developed countries e.g.
European countries, USA, UK etc.2There is no consensus regarding completeness of pass
through. Some studies reported completeness in the pass through process with respect to
bench mark monetary policy instrument e.g. Bernanke and Gertler (1995), Kashyap and
Stein (2000) and Althunbas et al. (2002) and Cook Steve (2008). Whereas, some studies
contradicted these findings and provided empirical evidence in favor of incompleteness
of pass through recently. The studies also found heterogeneity across countries, financial
institutions and retail bank products e.g. Mojon (2000), Bond t (2002), Hoffman and
Mizen (2004), Liu Ming-Hua et al. (2008) and Ozdemir Bilge Kagan (2009). The studies
focusing on pass through in Euro zone used Non-Harmonized Retail Interest Rate
Statistics (NRIR) e.g.Mojon (2000), Heinemann and Schuler (2002), Debondt(2002,
2005), Toolseema et.al. (2002), Sander and Kleimeir(2002, 2004). Some studies used
individual retail bank data, Cottarelli, Ferri and Generale(1995), Weth (2002),
Gambacorta(2004) and Degraeve Dejonghe and Vannet (2004).
Ozdemir Bilge Kagan (2009) estimated pass through between money market rate and
bank’s retail rate for Turkey. There is a limited body of literature on developing countries
like Pakistan. Qayyum, Khan, and Khawja (2005) estimated the pass through of TB rate
on Call Money Rate, saving deposit rate, six-month deposit rate and lending rate. They
used 6 month deposit rate and lending rate data from 1991:03 to 2004:12 and used

1
Romer and Romer (1989) and Bernanke Blinder (1992)
2
Mojon(2000), Bond t(2002), Graeve Ferre De et al. (2004), Kleimeier S. and Sander H. (2006), Sorensen
C. Kok et al. (2006), Liu Ming-Hua et al. (2008) etc.
Transfer Function Approach. Mohsin and Rivers (2011) measured pass through between
Treasury Bill (TB) rate and retail rates in the case of Pakistan. They found that although
pass through is not complete but the degree of pass through is moderately high. The State
Bank of Pakistan presently using Discount Rate as a monetary policy tool although in the
past has used various other tools too.
This study attempts to measure pass through between DR and retail rates using monthly
data from November 2001 to March 2011. The weighted average lending and deposit
rates data of four types of banks i.e. Private Domestic banks, Foreign Banks,
Nationalized banks and specialized banks using the weighted average lending and deposit
rate. Furthermore, the study intends to include a time dummy to capture the effect of
monetary policy restriction on the pass through rate. It will help to find whether banks
change their pass through rate while there is an increase in policy rate. The comparison of
pass through in the case of various categories of banks’ retail rates and DR will be an
interesting extension of the ongoing research in Pakistan.

2. Methodology: My study intends to use Panel data techniques. First I will apply Panel
Unit Root Tests to check for the stationarity of data. Then Pedroni Panel Cointegration
tests shall be applied to check for long run relationship between DR and Lending and
Deposit rate. Further, Phillips and Loretain (1991) method which is an extension of Engle
and Granger (1988) will be applied with cross section dummies.

2.1 Panel Unit Root:

This study will utilize three panel unit root tests to assess stationary data of lending,
deposit and DR (Levin et al., 2002; Im et al., 2003; Hadri, 2000).

The Levin, Lin and Chu (LLC) test assumes that persistence parameters remain the same
across cross sections. This means that ψi=ψ for all i. Alternatively, the Im, Pesaran and
Shin (IPS) test allows ψ to vary across all cross sections.

The LLC model allows for fixed effects and unit specific time trend along with common
time effects. The structure of their model is shown below:

yit   i   i t   t   i yi ,t 1   it , i=1,2…N,t=1,2…T (1)

The unit specific fixed effect is important to capture heterogeneity since the coefficient of
the lagged dependant variable is homogeneous across all cross sections in equation 1. Im
et al. (2003) extended the LLC framework by allowing heterogeneity in ρi under the
alternative hypothesis. The Lagrange Multiplier tests of Hadri (2000) have a different
null hypothesis than other panel unit root tests. The comparison of the results from all
three types of tests will be an interesting and will provide stronger evidence.3

2.2 Panel Cointegration:

This study attempts to use the Pedroni (1999, 2004) panel cointegration test to
estimate a long run relationship between bank and TB rates. Pedroni derived nearly seven
tests which are within- and between-dimensions. Pedroni’s tests are residual based,
similar to the Engle Granger tests. The slope coefficients vary over cross section units,
thereby allowing heterogeneity within the model. The following panel equation is
estimated:

(2)

Whereas, i =1, 2………N cross sectional units, t=1, 2………..T time periods, and
represents the column vector which consists of M independent variables for each ith
unit. M represents the number of independent variables. The variables Y and X are
considered to be non stationary, I (1) integrated of order one. The residual will be non
stationary, I(1), under the hypothesis of no cointegration. The parameters tend
to capture cross sectional fixed effects and deterministic trends, respectively. The
separate slope coefficients ensure that cointegrating vectors may also be heterogeneous.

In order to compute the required panel cointegrating statistic, Equation 1 is estimated by

OLS, for every individual cross section. The within-dimension based estimates are panel
and panel t statistics. They are derived by computing the first difference of all
variables.

(3)

3
For details see Mohsin and Rivers (2011) and Baner jee A (1999)
2.3 Phillips and Loretain (PL, 1991) Method

The marginal cost pricing model uses Engle Granger type equation4

(4)

i=1,2,…..,N and t=1,2………T

Where represents bank lending or deposit rate; , the monetary policy instrument,
DR, money market rate or federal fund rate; , the residual term; and measure the
mark up and long run degree of pass through respectively.

Liu et al. (2008) estimated the following triangular system of equations to model long run
relationship between policy rate and market rates;

, t=1, 2,……, T (5)

Xit =Xit-1+U2it (5a)

Where is a stationary vector?

The estimation of equation 1 requires both interest rates to be non stationary. If the
is not stationary then U2it interest rates will not cointegrate, thereby, resulting in a
spurious estimate.

Liu et al. reveals that even if U1it is stationary, OLS estimates of equations 1 and 1a do
not have standard distribution when U1it and U2it are correlated. Phillips and Loretain
(1991) suggested inclusion of leads and lags of the first difference in X t, . They
estimated the following equation:

+ (6)

The parameter estimates are unbiased asymptotically and normally distributed. Use of
this model provides two additional advantages. First, this model considers structural
changes, if they should occur. Second, it addresses past policy surprises and future policy
settings with regard to policy instrument and bank rates.

3.1 Discussion of Long Run Results

4
Rousseas (1985) and Bondt Gabe De (2002, 2005)
Table 1 provides the summary of Panel Unit Root tests (individual as well as common
process) which are applied to check the stationarity of three variables i.e. weighted
average lending and deposit rates and Discount rate (Monetary Policy Instrument). In
most of the cases I failed to reject the null hypothesis of unit root at levels since the
computed probabilities are greater than 0.05. But i can reject the hypothesis of unit root at
first difference in most of the cases which implies that the variables are non stationary at
levels but stationary at first difference, hence integrated of order one. In the case of Hadri
test I can reject the null hypothesis of stationarity in the case of all the three variables but
failed to reject the null hypothesis at first difference. Now in order to find evidence of
long run relationship, the study applies panel co integration analysis.
Table 2 summarizes the residual based Pedroni’s panel cointegration test between lending
rate and DR. In the within dimension test, seven out of eight tests reject the null
hypothesis of no cointegration. Similarly, in the between dimension case the null
hypothesis of no cointegration can be rejected for two out of three cases. The group ADF
stat has a probability of 0.06 which is higher than 0.05. Over all there is ample evidence
of long run relationship between lending rate and DR with Pedroni Cointegration test.
Now the study moves on check the long run relationship between Deposit rate and
Discount rate. Table 3 provides Pedroni’s residual based panel cointegration tests in the
case of deposit rate and DR. In eight within dimension tests, the computed probabilities
are estimated to be very high. I failed to reject the null hypothesis of no cointegration in
eight out of eight cases. Hence the case of no long run relationship between deposit rate
and DR is strong in the within dimension tests. Similarly, in the three between dimension
tests, the computed probabilities are more than 0.9 which are very higher than 0.05. I
failed to reject the null hypothesis of no cointegration in these tests too. Sorensen
Christopher Kok and Werner Thomas (2006) found similar results for saving deposits in
the Euro area using Pedroni residual based test. They are of the view that may be the
adjustment in saving deposit is so sluggish that there is no cointegration with the market
rate.
Table 1: Panel Unit Root Test

Method Null Hypothesis Lend Deposit DR

Stat Prob. Stat Prob. Stat Prob.

LLC- t* Unit Root (common
Stat Level process) -1.19 0.42 -3.03 0.99 -2.7 0.99
1st
Diff -15.84 0.00 -8.05 0.00 -6.87 0.00

HADRI-Z Level Stationary -6.7 0.00 9.5 0.00 8.67 0.00

1ST
Diff 2.64 0.06 -0.87 0.2 -0.73 0.77*

IPS-W Unit Root

Stat level (individual process) -1.28 0.10 3.2 0.99 -1.09 0.13
1st
Diff -16.57 0.00 -7.8 0.00 -10.24 0.00
Table 2: Pedroni Residual based Panel Cointegration Test
Variables: Lending rate and Discount Rate (DR).

Ho: No Cointegration

Weighted
Ha: with-in-dimension Statistic Prob Prob
Stat
Panel V Stat 1.9 0.03** 1.2 0.1***

Panel rho-Stat -1.8 0.04** -1.33 0.09***

Panel PP-Stat -1.9 0.02** -1.7 0.04**

Panel ADF-Stat -1.6 0.005** -1.5 0.0*

Ha: Between Dimension

Group rho-Stat -2.21 0.01*

Group PP-Stat -2.22 0.01*

Group ADF-Stat -1.5 0.06***

Note: *, ** and *** are significant at 1, 5 and 10 percent respectively.

3.2Phillips and Loretain (PL, 1991) estimates with slopes and intercepts dummies:

Table 4 reports 3 types of estimated equations for both lending and deposit rates.
Equation 1 estimates PL model without bank type dummies. It incorporates
autoregressive order 1,2 to tackle autocorrelation. The Equation 2 includes overall slope
and bank type dummies. Wald test has been applied to check whether bank types pass
through is different than the overall estimated slope. The Equation 3 provides slope as
well as intercept dummies but includes only Autoregressive scheme of order 1. Equation
1 for lending rate estimates the overall slope to be 0.20 which is far below one. It implies
that the pass through of DR with weighted average lending rate is not complete. It is
noted that the overall pass through parameter decreases with the inclusion of
autoregressive terms but Durbin Watson (DW) improves. It means banks pass on only
20% impact of a change in DR to lenders immediately which is very low. The overall
pass through in Eq 2 is estimated at 0.10 implying banks pass on only 10% impact of DR
to lenders in first month overall. The estimated slope dummies show that the pass through
is lowest in the case of Specialized banks and highest in the case of Domestic private
banks followed by foreign banks.

Table 3: Pedroni Residual Based Panel Cointegration Test

Variables: Deposit and Discount Rate (DR)

Ho: No Cointegration

Weighted
Ha: with-in-dimension Statistic Prob. Prob.
Stat
Panel V Stat 0.28 0.4 0.55 0.3

Panel rho-Stat 1.7 0.95 1.4 0.91

Panel PP-Stat 2.71 0.99 2.24 0.98

Panel ADF-Stat 2.61 0.99 2.13 0.98

Ha: Between Dimension

Group rho-Stat 1.3 0.89

Group PP-Stat 2.44 0.99

Group ADF-Stat 2.34 0.99

Note: None of computed probabilities less than 0.10, not significant.

The estimated pass through of private banks is 0.32 followed by foreign banks at 0.25.
The estimated pass through of nationalized banks is 0.19. I applied Wald test with the
coefficient restriction all slopes being equal. The estimated chi-sq value is 7146431 and
probability is zero. The null hypothesis of equality of slopes can be rejected in favor of at
least one slope coefficient is different. It implies there is heterogeneity in the response of
weighted average lending rates change when DR rate is changed as a tool to implement
monetary policy. The DW is 1.80 and goodness of fit explains 97% variations in the
model.
The Eq-3 of lending rate estimates overall pass through along with slope and intercept
dummies. The overall pass through is 0.15 but the estimated slope parameter of private
banks is o.32 followed by 0.24 for nationalized banks and 0.15 for foreign banks. The
equation includes only Autoregressive of degree one AR (1) scheme, the DW statistic is
1.3. After the inclusion of AR (2), the results are similar to Equation 2. The study
therefore considers Equation 2 as final result. The Eq 4 includes a slope dummy (DR-
Dum5) which is 1 after 2005-2010 and 0 otherwise. It is the time State Bank of Pakistan
pursued contractionary monetary policy ( a consistent rise in discount rate). This slope
dummy is positive and significant which implies that banks increased pass through during
a rise in monetary policy instrument.
The estimated pass through is low which implies that Pakistani banks only pass on
marginal impact of a change in discount rate to the lenders. It also implies that the
effectiveness of monetary policy is limited in the first month and takes time for the
complete impact.
Table 4: PL (1991) Estimates with dummy variables
Lending-DR
Eq-1 Eq-2 Eq-3 Eq-4
7.8 7.82 8.1 8.3
(9.5)* (9.8)* (17.3)* (10.8)*
DR (PL) 0.20 0.10 0.15 0.10
(2.7)** (1.8)*** (3.2)* (1.6)***
DR-FB 0.15 0.15 0.02
(39.7)* (43.9)* (11.9)*
DR-NB 0.09 0.24 0.004
(16.2)* (104.3)* (4.3)*
DR-PB 0.22 0.32 0.003
(37.5)* (138.2)* (4.9)*
DR-SB

DDR 0.25 0.25 0.2 0.92

(5.7)* [5.6]* (3.05)* (7.6)*
DERL 0.50 0.50 0.50 -0.8
(44.6)* (45.3)* (262.5)* (-106)*
C1-FB -2.4 -0.8
(-23.8)* (-2.6)*
C2-PB -3.37 0.91
(-79.1)* (4.3)*
C3-NB -1.8 1.3
(-24.7)* (4.7)*

DR*DUM5 0.003
(3.4)*
Chi-Sq 7146431*

R-sq 0.97 0.97 0.97 0.99

DW 1.8 1.8 1.3 1.76

Table 5: SR Analysis for Deposit and Discount Rate

Eq.1

D (DEP) = -0.002 + 0.16*D(DR)

(-0.13) (3.5)*
=0.05 F-Stat=3.8 Prob. (F-Stat) =0.002 DW=1.93

EQ. 2

D (DEP) = -0.02*D (DR) + 0.20*D (DNB) + 0.16*D (DPB) + 0.3*D (DFB)

(-0.4)* (1.92) ** (1.40) (1.96) **

=0.05 F-Stat=2.5 Prob. (F-Stat) =0.02 DW=2.3

EQ. 3

D (DEP) = -0.003*D(DR) + 0.20*D(DNB) + 0.2*D(DPB) + 0.3*D(DFB) + 0.02*D5(DR

(-0.04) (1.9) *** (1.3) (2.01) ** (0.2)

AR (1) =-0.14]
(-2.04)*

=0.07 F-Stat=2.98 Prob. (F-Stat) =0.003 DW=1.94

Note: *, **, *** imply significant at 1, 5 and 10 % respectively

4. SR Analysis for Deposit and Discount Rate:

Since deposit rate and discount rate do not have long run relationship, the study extends
the analysis to short run. Table 5 reports the results in the form of three equations. The
variables are in first difference form. The Eq 1 estimates fixed effect model where there
is overall pass through of discount and deposit rate is estimated in short run. The pass
through is 0.16 and it is found statistically significant. It implies that overall banks pass
on 16 % impact of a change in discount rate in the first month to depositors.
The Equation 2 adds banks type wise slope dummies. The overall slope is zero but the
pass through of nationalized and foreign banks is found to be 0.2 and 0.3 respectively.
The parameter pertaining to privatized banks is 0.16 but is not statistically significant.
Furthermore, goodness of fit measure is only 0.05.
Then equation 3 estimates the same model with AR (1) since the DW statistic is 2.3 in
Eq 2. The overall parameter is not statistically significant from zero where as the slope
parameters pertaining to nationalized and foreign banks are 0.2 and 0.3 respectively. The
goodness of fit measure improves slightly to 0.07 and DW statistic is estimated at 1.94.
The study finds overall pass through of 0.16 in the short run between deposit and
discount rate. There is found evidence of asymmetry since the pass through of various
bank types is different. Overall the higher pass through of 0.3 is estimated for foreign and
0.2 for the nationalized banks. The pass through of privatized banks is estimated at 0.20
but it is not significant.

5. Conclusion

This study estimated impact of monetary policy instrument, discount rate on weighted
average lending and deposit rates in Pakistan. The study has used bank type monthly data
from November 2001 to March 2011. The four bank types are nationalized, privatized,
foreign and specialized banks. Panel data techniques have been used for estimation of
results.
The study finds that all the three variables are non stationary at levels and stationary at
first difference. Pedroni Panel Cointegration technique has been applied for estimating
long run relationship which reveals that lending rate is cointegrated while deposit rate is
not cointegrated with the monetary policy instrument. Since lending and DR are
cointegrated, PL (1991) method has been applied with bank type dummies. The results
show that the overall lending rates pass through is 0.20 which is very low. However there
is evidence of asymmetry among various bank types since the pass through of private,
foreign and nationalized banks is estimated at 0.32, 0.25 and 0.19 respectively. The
study finds evidence that lending rates pass through increased significantly during
monetary policy restriction times but deposit rate remains same. The short run analysis in
the case of deposit and discount rate show overall banks pass on 16 percent impact of
discount rate to depositors. The pass through of Foreign and nationalized banks is 0.30
and 0.20 respectively. It further clarifies the idea that banks’ profitability may be at the
cost of low deposit rates and depositors are not passed on the marginal benefit of a rise in
discount rate during restricted monetary policy times.
The estimates suggest that the overall effectiveness of monetary policy is limited and
there is significant lag in its completeness.
References

Baltagi Badi H. (2000) “Non stationary panels, panel cointegration and dynamic panels:
Advances in Econometrics, Vol. 15. ELSEVIER SCIENCE inc.

Bernanke, B., and Alan Blinder (1992) “Credit, Money, and Aggregate Demand”
American Economic Review 82, 901–21

Bernanke, B., and Mark Gertler (1995) Inside the Black Box: The Credit Channel of
Monetary Policy Transmission. Journal of Economic Perspectives 9:4, 27–48.

Bondt, G. (2002) Retail Bank Interest Rate Pass-through: New Evidence at the Euro Area
Level. Frankfurt: European Central Bank. (ECB Working Paper No. 136.)

Cottarelli, C., and Angeliki Kourelis (1994) “Financial Structure, Bank Lending Rates,
and the Transmission Mechanism of Monetary Policy” IMF Staff Papers

41:4, 587–623.

Gropp Kok Sørensen, and Lichtenberger(2007) " The Dynamics of Bank spreads and
Financial Structure." European Central Bank, Working paper series no. 714.

Hanan, T., and Allen Berger (1991) The Rigidity of Prices: Evidence from the Banking
Industry. American Economic Review 81:4, 938–945.

Mishkin, Frederic S. (1995) Symposium on the Monetary Transmission

Mechanism. Journal of Economic Perspectives 9:4, 3–10.

Hofmann, B., Mizen, P., 2004. Interest rate pass-through and monetary transmission:
Evidence from individual financial institutions’ retail rates. Economica 71, 99–123.

ulsewig Oliver et al (2006) “Bank loan supply and monetary policy transmission in
Germany: An assessment based on matching impulse responses.” Journal of Banking and
Finance 30 (2006) 2893-2910.

Kashyap, A.K., Stein, J., 2000. What do a million observations on banks say about the
transmission of monetary policy? American Economic Review 86, 310–314.
Kleimeier, S., Sander, H., 2006. Expected versus unexpected monetary policy impulses
and interest rate pass through in euro zone’s retail banking markets. Journal of Banking
and Finance30, 1839–1870.

Leuvensteijn Michiel van et al (2008) “Impact of bank competition on the interest rate
pass through in the Euro area.” European Central Bank, Working Paper Series NO 885 /
MARCH 2008

Leuven and Leuven (2001)” Bank lending rate pass through and differences in
transmission of a single EMU monetary policy.” Katholieke University, Center for
Economic studies discussion paper series (DPS)01.17

Levin, Lin and Chu (LLC, 2004)” Unit root tests in panel data: asymptotic and finite
sample properties.” Journal of Econometrics 108 (2002) 1-24

Liu M.H et al (2008) “Monetary Policy transparency and pass through of retail interest
rates” Journal of Banking & Finance 32 (2008) 501–511

Mohsin H. and Rivers P. (2011) “Are Domestic Banks’ Interest Rate Pass through Higher
than Foreign Banks? Empirical Evidence from Pakistan” International Journal of
Economics and Finance, November 2011
Mojon, B. (2000) “Financial Structure and the Interest Rate Channel of ECB Monetary
Policy” Frankfurt: European Central Bank (ECB Working Paper No. 40)

Orsal Deniz Dilan Karaman (2007)” Comparison of Panel Cointegration tests”. SBF 649
Discussion Paper 2007-029 Humboldt-Universität zu Berlin, Germany

Pedroni Peter (2004)” Panel Cointegration: Asymptotic and Finite sample properties of
pooled time series with an application to the PPP hypothesis”. Econometric Theory 20,
2004, 597-625

Pedroni Peter (1999)” Critical values for cointegration tests in heterogeneous panels with
multiple regressors.” Oxford Bulletin of Economics and Statistics, special issue (1999)

Phillips, P.C.B., Loretain, M., 1991. Estimating long-run equilibria. The Review of
Economic Studies 58, 407–436
Qayyum, Khan and Khawaja (2005) " Interest rate pass through in Pakistan: Evidence
from Transfer Function Approach. The Pakistan Development Review 44: 4 Part II
(Winter 2005)

Romer, C., and D. Romer (1989) Does Monetary Policy Matter? A New Test in the Spirit
of Friedman and Schwartz. NBER Macroeconomic Annual 4, 121–170

Sander and Kleimeir (2004)”Interest rate pass through in an enlarged Europe: the role of
banking market structure for monetary policy transmission in transitional countries.”
Unpublished Mimeo

Tieman Alexander (2004)” Interest rate pass through in Romania and other Central
European economies”. IMF Working Paper WP/04/211.

Vega, E., and Alessandro Rebucci (2003) Retail Bank Interest Rate Pass-through: Is
Chile Atypical? International Monetary Fund (IMF Working Paper (WP/03/112)